[PATCH] S4bios for 2.5.52.

[PATCH] S4bios for 2.5.52.

[Ducrot Bruno]

[-- Attachment #1: Type: text/plain, Size: 567 bytes --]

This patch add s4bios support for 2.5.52.

S4bios is an alternative for the ACPI S4 system suspend state, but is a bit
more easy to implement.  It suppose though that the BIOS support this feature.
For some BIOS, creating a so-called suspend partition with the help
of lphdisk is OK.

Plus, it permit for Pavel to have a nice graphic display at suspend/resume. 

echo 4 > /proc/acpi/sleep is for swsusp;
echo 4b > /proc/acpi/sleep is for s4bios.

Cheers,

-- 
Ducrot Bruno
http://www.poupinou.org        Page profaissionelle
http://toto.tu-me-saoules.com  Haume page

[-- Attachment #2: s4bios-2.5.52.diff --]
[-- Type: text/plain, Size: 7973 bytes --]

 arch/i386/kernel/acpi_wakeup.S  |   33 +++++++++++++++++++--
 drivers/acpi/acpi_ksyms.c       |    1 
 drivers/acpi/hardware/hwsleep.c |   61 ++++++++++++++++++++++++++++++++++++----
 drivers/acpi/include/acpixf.h   |    4 ++
 drivers/acpi/sleep.c            |   28 +++++++++++++++---
 include/linux/suspend.h         |    2 +
 6 files changed, 117 insertions(+), 12 deletions(-)

arch/i386/kernel/acpi_wakeup.S
--- linux-2.5.52/arch/i386/kernel/acpi_wakeup.S	2002/12/17 19:15:12	1.1
+++ linux-2.5.52/arch/i386/kernel/acpi_wakeup.S	2002/12/17 19:20:34
@@ -272,7 +272,7 @@
 
 ENTRY(do_suspend_lowlevel)
 	cmpl $0,4(%esp)
-	jne .L1432
+	jne ret_point
 	call save_processor_state
 
 	movl %esp, saved_context_esp
@@ -285,7 +285,7 @@
 	movl %edi, saved_context_edi
 	pushfl ; popl saved_context_eflags
 
-	movl $.L1432,saved_eip
+	movl $ret_point,saved_eip
 	movl %esp,saved_esp
 	movl %ebp,saved_ebp
 	movl %ebx,saved_ebx
@@ -297,7 +297,7 @@
 	addl $4,%esp
 	ret
 	.p2align 4,,7
-.L1432:
+ret_point:
 	movl $__KERNEL_DS,%eax
 	movw %ax, %ds
 	movl saved_context_esp, %esp
@@ -310,6 +310,33 @@
 	movl saved_context_edi, %edi
 	call restore_processor_state
 	pushl saved_context_eflags ; popfl
+	ret
+
+ENTRY(do_suspend_lowlevel_s4bios)
+	cmpl $0,4(%esp)
+	jne ret_point
+	call save_processor_state
+
+	movl %esp, saved_context_esp
+	movl %eax, saved_context_eax
+	movl %ebx, saved_context_ebx
+	movl %ecx, saved_context_ecx
+	movl %edx, saved_context_edx
+	movl %ebp, saved_context_ebp
+	movl %esi, saved_context_esi
+	movl %edi, saved_context_edi
+	pushfl ; popl saved_context_eflags
+
+	movl $ret_point,saved_eip
+	movl %esp,saved_esp
+	movl %ebp,saved_ebp
+	movl %ebx,saved_ebx
+	movl %edi,saved_edi
+	movl %esi,saved_esi
+
+	pushl $3
+	call acpi_enter_sleep_state_s4bios
+	addl $4,%esp
 	ret
 
 ALIGN
drivers/acpi/include/acpixf.h
--- linux-2.5.52/drivers/acpi/include/acpixf.h	2002/12/17 19:15:16	1.1
+++ linux-2.5.52/drivers/acpi/include/acpixf.h	2002/12/17 19:17:11
@@ -380,6 +380,10 @@
 	u8                      sleep_state);
 
 acpi_status
+acpi_enter_sleep_state_s4bios (
+	void);
+
+acpi_status
 acpi_leave_sleep_state (
 	u8                      sleep_state);
 
drivers/acpi/hardware/hwsleep.c
--- linux-2.5.52/drivers/acpi/hardware/hwsleep.c	2002/12/17 19:15:17	1.1
+++ linux-2.5.52/drivers/acpi/hardware/hwsleep.c	2002/12/17 19:39:30
@@ -184,6 +184,7 @@
 acpi_enter_sleep_state (
 	u8                      sleep_state)
 {
+	int                     i;
 	u32                     PM1Acontrol;
 	u32                     PM1Bcontrol;
 	ACPI_BIT_REGISTER_INFO  *sleep_type_reg_info;
@@ -298,16 +299,18 @@
 	}
 
 	/* Wait until we enter sleep state */
+	/* But gracefully fail if we wait too long. */
 
+	i = 10;
 	do {
 		status = acpi_get_register (ACPI_BITREG_WAKE_STATUS, &in_value, ACPI_MTX_LOCK);
 		if (ACPI_FAILURE (status)) {
 			return_ACPI_STATUS (status);
 		}
-
+		acpi_os_stall (1000);
 		/* Spin until we wake */
 
-	} while (!in_value);
+	} while (!in_value && --i);
 
 	status = acpi_set_register (ACPI_BITREG_ARB_DISABLE, 0, ACPI_MTX_LOCK);
 	if (ACPI_FAILURE (status)) {
@@ -319,6 +322,54 @@
 
 /******************************************************************************
  *
+ * FUNCTION:    Acpi_enter_sleep_state_s4bios
+ *
+ * PARAMETERS:  None
+ *
+ * RETURN:      Status
+ *
+ * DESCRIPTION: Perform a s4 bios request.
+ *              THIS FUNCTION MUST BE CALLED WITH INTERRUPTS DISABLED
+ *
+ ******************************************************************************/
+
+acpi_status
+acpi_enter_sleep_state_s4bios (void)
+{
+	u32                     in_value;
+	int                     i;
+	acpi_status             status;
+
+
+	ACPI_FUNCTION_TRACE ("Acpi_enter_sleep_state_s4bios");
+
+
+	acpi_set_register (ACPI_BITREG_WAKE_STATUS, 1, ACPI_MTX_LOCK);
+	acpi_hw_clear_acpi_status();
+
+	acpi_hw_disable_non_wakeup_gpes();
+
+	ACPI_FLUSH_CPU_CACHE();
+
+	acpi_os_stall (10000);
+
+	status = acpi_os_write_port (acpi_gbl_FADT->smi_cmd, (acpi_integer) acpi_gbl_FADT->S4bios_req, 8);
+
+	i = 10;
+	do {
+		status = acpi_get_register (ACPI_BITREG_WAKE_STATUS, &in_value, ACPI_MTX_LOCK);
+		if (ACPI_FAILURE (status)) {
+			return_ACPI_STATUS (status);
+		}
+		acpi_os_stall (1000);
+	} while (!in_value && --i);
+
+	return_ACPI_STATUS (AE_OK);
+}
+
+
+/******************************************************************************
+ *
  * FUNCTION:    Acpi_leave_sleep_state
  *
  * PARAMETERS:  Sleep_state         - Which sleep state we just exited
@@ -341,6 +392,9 @@
 	ACPI_FUNCTION_TRACE ("Acpi_leave_sleep_state");
 
 
+	/* Be sure to have BM arbitration */
+	status = acpi_set_register (ACPI_BITREG_ARB_DISABLE, 0, ACPI_MTX_LOCK);
+
 	/* Ensure Enter_sleep_state_prep -> Enter_sleep_state ordering */
 
 	acpi_gbl_sleep_type_a = ACPI_SLEEP_TYPE_INVALID;
@@ -371,9 +425,6 @@
 	if (ACPI_FAILURE (status)) {
 		return_ACPI_STATUS (status);
 	}
-
-	/* Disable BM arbitration */
-	status = acpi_set_register (ACPI_BITREG_ARB_DISABLE, 0, ACPI_MTX_LOCK);
 
 	return_ACPI_STATUS (status);
 }
drivers/acpi/sleep.c
--- linux-2.5.52/drivers/acpi/sleep.c	2002/12/17 19:15:18	1.1
+++ linux-2.5.52/drivers/acpi/sleep.c	2002/12/17 19:47:28
@@ -149,8 +149,15 @@
 		if (state > ACPI_STATE_S1) {
 			error = acpi_save_state_mem();
 
+			/*
+			 * Do not kill this #if 0 please.
+			 * If one day we get S4, we should be happy to remember
+			 * where to save to disk.
+			 */
+#if 0
 			if (!error && (state == ACPI_STATE_S4))
 				error = acpi_save_state_disk();
+#endif
 
 			if (error) {
 				device_resume(RESUME_RESTORE_STATE);
@@ -227,6 +234,8 @@
 	case ACPI_STATE_S3:
 		do_suspend_lowlevel(0);
 #endif
+	case ACPI_STATE_S4:
+		do_suspend_lowlevel_s4bios(0);
 		break;
 	}
 	local_irq_restore(flags);
@@ -250,10 +259,14 @@
 	if (state < ACPI_STATE_S1 || state > ACPI_STATE_S5)
 		return AE_ERROR;
 
+	/* Since we handle S4 via a different path (swsusp), give up if no s4bios. */
+	if (state == ACPI_STATE_S4 && !acpi_gbl_FACS->S4bios_f)
+		return AE_ERROR;
+
 	freeze_processes();		/* device_suspend needs processes to be stopped */
 
 	/* do we have a wakeup address for S2 and S3? */
-	if (state == ACPI_STATE_S2 || state == ACPI_STATE_S3) {
+	if (state == ACPI_STATE_S2 || state == ACPI_STATE_S3 || state == ACPI_STATE_S4) {
 		if (!acpi_wakeup_address)
 			return AE_ERROR;
 		acpi_set_firmware_waking_vector((ACPI_PHYSICAL_ADDRESS) acpi_wakeup_address);
@@ -296,8 +309,11 @@
 	ACPI_FUNCTION_TRACE("acpi_system_sleep_seq_show");
 
 	for (i = 0; i <= ACPI_STATE_S5; i++) {
-		if (sleep_states[i])
+		if (sleep_states[i]) {
 			seq_printf(seq,"S%d ", i);
+			if (i == ACPI_STATE_S4 && acpi_gbl_FACS->S4bios_f)
+				seq_printf(seq, "S4b ");
+		}
 	}
 
 	seq_puts(seq, "\n");
@@ -336,12 +352,12 @@
 	if (!sleep_states[state])
 		return_VALUE(-ENODEV);
 
+	if (state == 4 && state_string[1] != 'b') {
 #ifdef CONFIG_SOFTWARE_SUSPEND
-	if (state == 4) {
 		software_suspend();
 		return_VALUE(count);
-	}
 #endif
+	}
 	status = acpi_suspend(state);
 
 	if (ACPI_FAILURE(status))
@@ -660,6 +676,10 @@
 		if (ACPI_SUCCESS(status)) {
 			sleep_states[i] = 1;
 			printk(" S%d", i);
+		}
+		if (i == ACPI_STATE_S4 && acpi_gbl_FACS->S4bios_f) {
+			sleep_states[i] = 1;
+			printk(" S4b");
 		}
 	}
 	printk(")\n");
drivers/acpi/acpi_ksyms.c
--- linux-2.5.52/drivers/acpi/acpi_ksyms.c	2002/12/17 19:15:19	1.1
+++ linux-2.5.52/drivers/acpi/acpi_ksyms.c	2002/12/17 19:17:11
@@ -86,6 +86,7 @@
 EXPORT_SYMBOL(acpi_get_register);
 EXPORT_SYMBOL(acpi_set_register);
 EXPORT_SYMBOL(acpi_enter_sleep_state);
+EXPORT_SYMBOL(acpi_enter_sleep_state_s4bios);
 EXPORT_SYMBOL(acpi_get_system_info);
 EXPORT_SYMBOL(acpi_get_devices);
 
include/linux/suspend.h
--- linux-2.5.52/include/linux/suspend.h	2002/12/17 19:15:19	1.1
+++ linux-2.5.52/include/linux/suspend.h	2002/12/17 19:17:11
@@ -71,6 +71,8 @@
 
 extern void do_suspend_lowlevel(int resume);
 
+extern void do_suspend_lowlevel_s4bios(int resume);
+
 #else
 static inline void software_suspend(void)
 {

RE: [PATCH] S4bios for 2.5.52.

[Grover, Andrew]

> From: Ducrot Bruno [mailto:poup-kk6yZipjEM5g9hUCZPvPmw@public.gmane.org] 
> This patch add s4bios support for 2.5.52.
> 
> S4bios is an alternative for the ACPI S4 system suspend 
> state, but is a bit
> more easy to implement.  It suppose though that the BIOS 
> support this feature.
> For some BIOS, creating a so-called suspend partition with the help
> of lphdisk is OK.
> 
> Plus, it permit for Pavel to have a nice graphic display at 
> suspend/resume. 
> 
> echo 4 > /proc/acpi/sleep is for swsusp;
> echo 4b > /proc/acpi/sleep is for s4bios.

I still am not clear on why we would want s4bios in 2.5.x, since we have S4.
Like you said, S4bios is easier to implement, but since Pavel has done much
of the heavy lifting required for S4 proper, I don't see the need.

Regards -- Andy

RE: [PATCH] S4bios for 2.5.52.

[Kai Germaschewski]

On Tue, 17 Dec 2002, Grover, Andrew wrote:

> > From: Ducrot Bruno [mailto:poup-kk6yZipjEM5g9hUCZPvPmw@public.gmane.org] 
> > This patch add s4bios support for 2.5.52.

> > echo 4 > /proc/acpi/sleep is for swsusp;
> > echo 4b > /proc/acpi/sleep is for s4bios.
> 
> I still am not clear on why we would want s4bios in 2.5.x, since we have S4.
> Like you said, S4bios is easier to implement, but since Pavel has done much
> of the heavy lifting required for S4 proper, I don't see the need.

Let me counter this, I have to admit that I didn't try the patch yet, but
my laptop does S4 BIOS, and I'd definitely prefer that to swsusp, since
it's much faster and also I somewhat have more faith into S4 BIOS than
swsusp (as in: after resuming, it'll most likely either work or crash, but
not cause any weird kinds of corruption). Since it does not need not much
more to support it than S3, I don't see why you wouldn't want to give
users the option?

--kai

RE: [PATCH] S4bios for 2.5.52.

[Grover, Andrew]

> From: Kai Germaschewski [mailto:kai-K3FV8rsiL1oTdZ4mcS17cL1bcanuHkqH@public.gmane.org] 
> > I still am not clear on why we would want s4bios in 2.5.x, 
> since we have S4.
> > Like you said, S4bios is easier to implement, but since 
> Pavel has done much
> > of the heavy lifting required for S4 proper, I don't see the need.
> 
> Let me counter this, I have to admit that I didn't try the 
> patch yet, but
> my laptop does S4 BIOS, and I'd definitely prefer that to 
> swsusp, since
> it's much faster and also I somewhat have more faith into S4 BIOS than
> swsusp (as in: after resuming, it'll most likely either work 
> or crash, but
> not cause any weird kinds of corruption). Since it does not 
> need not much
> more to support it than S3, I don't see why you wouldn't want to give
> users the option?

Ok that's reasonable.

My belief is that S4bios is a stopgap measure until S4 gets better. That
said, I think you are right - it should go in for now, and then at some
point in the future someone will say, "S4bios?? who needs *that* anymore??"
and it will get pulled out. ;-)

So I'll look to merge it, unless someone upstream beats me to it.

Regards -- Andy

Re: [PATCH] S4bios for 2.5.52.

[Pavel Machek]

Hi!

> > > I still am not clear on why we would want s4bios in 2.5.x, 
> > since we have S4.
> > > Like you said, S4bios is easier to implement, but since 
> > Pavel has done much
> > > of the heavy lifting required for S4 proper, I don't see the need.
> > 
> > Let me counter this, I have to admit that I didn't try the 
> > patch yet, but
> > my laptop does S4 BIOS, and I'd definitely prefer that to 
> > swsusp, since
> > it's much faster and also I somewhat have more faith into S4 BIOS than
> > swsusp (as in: after resuming, it'll most likely either work 
> > or crash, but
> > not cause any weird kinds of corruption). Since it does not 
> > need not much
> > more to support it than S3, I don't see why you wouldn't want to give
> > users the option?
> 
> Ok that's reasonable.
> 
> My belief is that S4bios is a stopgap measure until S4 gets better. That
> said, I think you are right - it should go in for now, and then at some
> point in the future someone will say, "S4bios?? who needs *that* anymore??"
> and it will get pulled out. ;-)

Thanx. I'm looking forward to see it merged.
								Pavel
-- 
Casualities in World Trade Center: ~3k dead inside the building,
cryptography in U.S.A. and free speech in Czech Republic.


-------------------------------------------------------
This sf.net email is sponsored by:
With Great Power, Comes Great Responsibility 
Learn to use your power at OSDN's High Performance Computing Channel
http://hpc.devchannel.org/

Re: [PATCH] S4bios for 2.5.52.

[Ducrot Bruno]

On Tue, Dec 17, 2002 at 04:14:41PM -0600, Kai Germaschewski wrote:
> On Tue, 17 Dec 2002, Grover, Andrew wrote:
> 
> > > From: Ducrot Bruno [mailto:poup-kk6yZipjEM5g9hUCZPvPmw@public.gmane.org] 
> > > This patch add s4bios support for 2.5.52.
> 
> > > echo 4 > /proc/acpi/sleep is for swsusp;
> > > echo 4b > /proc/acpi/sleep is for s4bios.
> > 
> > I still am not clear on why we would want s4bios in 2.5.x, since we have S4.
> > Like you said, S4bios is easier to implement, but since Pavel has done much
> > of the heavy lifting required for S4 proper, I don't see the need.
> 
> Let me counter this, I have to admit that I didn't try the patch yet, but
> my laptop does S4 BIOS, and I'd definitely prefer that to swsusp, since
> it's much faster and also I somewhat have more faith into S4 BIOS than
> swsusp (as in: after resuming, it'll most likely either work or crash, but
> not cause any weird kinds of corruption). Since it does not need not much
> more to support it than S3, I don't see why you wouldn't want to give
> users the option?
> 

In fact, I agree with Andrew.  But Pavel insist and want it,
and since it is really straightforward to have it in 2.5
because Pavel have already done all the work, why not giving
him a little gift for Christmas ?

A final word:  S4bios is way much _slower_ by nature than S4OS, especially
at suspend path, which is what we don't want if we have to go to S4 due
to an emergency.

Kai, if you see that S4OS is slower, it is probably a bug that have to be
fixed in swsusp.

Cheers,

-- 
Ducrot Bruno
http://www.poupinou.org        Page profaissionelle
http://toto.tu-me-saoules.com  Haume page

RE: [ACPI] Re: [PATCH] S4bios for 2.5.52.

[Nigel Cunningham]

> Kai, if you see that S4OS is slower, it is probably a bug
> that have to be
> fixed in swsusp.

There does seem to be a problem with the rw_swap_page_sync routine. Perhaps
not on Pavel's machine (:>), but I've experienced painfully slow suspends
too. Performance was helped immensely by applying a concept that I'll credit
Lyle Seaman for - unrolling rw_swap_page_sync and submitting all the IO for
pages then looping again and waiting on each sync. I found with writing 7000
or so pages (a suspend from init S) that something gets triggered to force
the sync about every 3900 pages anyway.

Regards,

Nigel